The second factor is increasing the significance of native growth compared to uber-stacks of angels, vivs, or demons. Because leveling up these uber-stacks can be quite difficult—they are constantly replenished, and they are too high level for efficient grinding.
And it's simply one more parameter characterizing development. The more such parameters there are, the more varied the development becomes.
Added after 34 minutes
The percentage of losses correlates most strongly with skill, not with settings.
Added after 12 minutes
By the way, regarding the question of distribution balance, a real-life example—a Castle vs Castle game. One player is given a couple of trashy Griffin Conservatorys, the other 2 external dwellings of pokemons and 1 of griffins.
If they played without experience, events would unfold like this: the first takes the Griffin Conservatorys (losing all their same-tier units there), breaks the Dragon Utopia on the 2nd week with angels, while the second just sits around waiting for the 3rd week.
But with experience, it happened differently—the first player's development was the same, but the second quickly leveled up the pokemons to rank 10, resulting in something like 120 very angry comrades, which in terms of PPS (penetration power parity) were equivalent to ~230+ basic pokemons, with which (supported by others) they also broke the Dragon Utopia on the 2nd.